The impact of changes in the weather on the surface temperatures of
Windermere (UK) and Lough Feeagh (Ireland)
GLEN GEORGE1, DIANE HEWITT2, ELEANOR JENNINGS3, NORMAN
ALLOTT3 AND PHILIP MCGINNITY4
1
CEH Lancaster Environment Centre, Library Avenue, Bailrigg, Lancaster,
LA1 4AP, UK
[email protected]
2
Freshwater Biological Association, The Ferry House, Far Sawrey, Ambleside, Cumbria, LA22 OLP,
UK
3
Trinity College, College Green, Dublin 2, Ireland
4
Marine Institute, Newport, County Mayo, Ireland
Abstract The surface temperatures of Lake Windermere (UK) and Lough Feeagh
(Ireland) have been recorded every day since 1960. Here, we examine the factors
influencing these measurements and their associated weekly variance. At both sites,
there was a progressive increase in the measured temperature but the rate of increase
was very much greater in Lake Windermere. The variance of the Lake Windermere
temperatures was negatively correlated with the cloud cover but there was no
corresponding change in Lough Feeagh. Comparisons with the Lamb system of
weather classification showed that the lake temperatures were closely correlated with
these synoptic conditions. The highest winter temperatures were recorded during
‘westerly’ conditions and the highest summer temperatures under ‘southerly’
conditions. The most striking difference between the lakes was their response to cold
winters and windy summers. Such results demonstrate that, lakes that are physically
very similar can still ‘filter’ the climate signal in subtly different ways
.
Key words Lake temperatures, weather patterns, climate change, UK, Ireland.
INTRODUCTION
Year-to-year changes in the weather have a major effect on the thermal characteristics
of lakes (George, 1989; George et al., 2004). Changes in the air temperature and
cloud cover regulate the flux of heat across the air-water interface whilst variations in
the wind speed influence the vertical dispersion of heat.
Lakes thus behave as
integrators of the local weather and can even amplify very weak meteorological
signals (George & Harris, 1985; George & Taylor, 1995). Very little attention has
hitherto been paid to the way in which lakes respond to short-term changes in the
weather. Climatologists have devised a number of different ways to quantify these
changes. Some, such as the North Atlantic Oscillation Index (Hurrell,1995), are
1
based on regional-scale gradients in the atmospheric pressure. Others, such as those
devised Lamb (1972), are based on the detailed analysis of daily weather maps. Here,
we use the system devised by Lamb for the British Isles to explore the impact of longterm changes in the weather on the surface temperatures of two deep, thermally
stratified lakes.
There are very few long-term records of the day-to-day variations in the surface
temperature of a lake. Most long-term studies rely on measurements taken at weekly
or fortnightly intervals. Several automatic monitoring stations have recently been
established in Europe (Rouen et al., 2000) but these records are too short for
systematic analysis.
In this paper, we analyze some long-term records of the
variations in the surface temperature of Lake Windermere in the UK and Lough
Feeagh in Ireland. Both lakes are situated on the Atlantic coast and the measurements
cover a 35-40 year period that extends from the early1960’s to the late 1990’s.
DESCRIPTION OF SITES AND METHODS
Windermere, the largest lake in the English Lake District, covers an area of 8.1 km2
and has a maximum depth of 60 m. It lies in a relatively sheltered valley (54º 22’ N;
2º 56’ W) and is divided into two basins by a large island. The surface temperatures
were measured in a sheltered bay using a mercury thermometer with a resolution of
0.1oC. All measurements were taken in early morning (ca 8.30 hours GMT) to
minimise the effects of near-surface heating. Lough Feeagh is a trough-like basin
situated on the shores of Clew Bay in the west of Ireland (53o 50’N; 9o35’W). It has
a surface area of 3.9 km2 and a maximum depth of 45 m. The surface temperatures
were recorded by a Negretti system installed on the shore near the main outflow. The
readings used were recorded at 06.00 GMT and the nominal resolution of the chart
2
was 0. 1oC. The raw data from both sites was organised into a days vs years matrix
and averages and standard deviations calculated for each week in each year. There
were few breaks in the records, but weekend measurements at Lake Windermere were
discontinued in the 1990’s. Where measurements were not available for the full
week, the means and standard deviations were calculated for the remaining 5 days.
The variations in the local weather
At LakeWindermere, meteorological data were acquired from two weather stations in
the village of Ambleside (54º 24’ N; 2º 57’ W). The first station was in operation
from 1960 to 1974 and the second from1971 to 2000. Here, we have combined the
results to form a harmonised time-series that extends from January 1960 to December
2000. The average air temperature was calculated from the daily minimum and
maximum and the cloud cover estimated by visual observation. At Lough Feeagh,
meteorological data was acquired from a station located about 50 km away near the
town of Belmullet (54º 13’ N; 10º 0’ W).Here all measurements were taken at hourly
intervals by an observer stationed on site.
The variations in the regional weather types
The Lamb classification was originally based on a subjective analysis of daily weather
charts but is now based on a multivariate analysis of the pressure gradients around the
British Isles (Jones et al., 1993). The classification contains eight directional types,
two non-directional types and an unclassified category. The directional types (N, NE,
E,SE, S, SW, W and NW) are defined according to the general direction of the air
flow. The non-directional types (A – anticyclonic and C – cyclonic) occur when high
3
or low pressures systems are dominant. Table 1 shows the seasonal frequencies of the
dominant circulation types as listed by Kelly et al. (1997).
Table 1 The frequency (%) of the dominant Lamb weather types between 1861 and 1990.
From Kelly et al. 1997.
Lamb Weather Type
Winter
Spring
Summer
Autumn
Westerly (W)
23.0
12.3
17.9
19.7
Anticyclonic (A)
16.1
17.3
18.9
18.9
Cyclonic (C)
10.7
12.0
16.1
12.1
Northerly (N)
3.3
5.6
4.8
4.6
The main circulation types are the westerly, the anticyclonic, the cyclonic and the
northerly.
There are no major differences in the seasonal distribution of these
categories but westerly conditions are more common in winter. Here, we relate the
seasonal variations in the lake temperatures and their variability to the changing
frequency of westerly and anti-cyclonic conditions. Records of the week-to-week
variation in these weather types were abstracted from a web page maintained by the
Climate Research Unit (www.cru.uea.ac.uk) and then correlated with the lake
measurements. In this way, we could relate both the seasonal variation in the surface
temperature of the lakes and their weekly variance to the changing frequency of the
dominant weather types.
4
RESULTS
The seasonal variation in the surface temperature
Fig. 1 shows the average week-to-week variation in the surface temperature of Lake
Windermere and Lough Feeagh.
Fig. 1 The seasonal variation in the average surface temperature of Lake
Windermere (-) and Lough Feeagh (--).
The seasonal variations recorded at the two sites were very similar but the summer
temperatures in Lake Windermere were about 2oC higher than that those in Lough
Feeagh. The main factor influencing the surface temperature of a deep lake is the
duration and intensity of thermal of stratification. The onset of stratification was
practically the same at the two sites but the autumn overturn was a few weeks earlier
in Lough Feeagh. The relative exposure of the two sites also has an effect on the
stability of the thermocline. The average depth of the thermocline in the two lakes is
about 13m but the temperature gradients that develop in Lake Windermere are rather
steeper than those reported from Lough Feeagh.
The long-term change in the surface temperature
Figs. 2a and 2b shows the long-term change in the surface temperature of Lake
Windermere and Lough Feeagh. The points show the annual means, the lines the
5
three-point running means and the linear regressions the trend. In both lakes, there
was a sustained increase in the average temperature but the rate of increase was
greater in Lake Windermere than in Lough Feeagh.
At Lake Windermere, the temperature increased by 1.4oC between 1960 and 2000.
At Lough Feeagh, the increase was only 0.7oC between 1960 and 1997. The other
difference between the lakes was the pattern of change experienced in the 1990’s. In
Lake Windermere, a rising trend was still evident but the temperature variations in
Lough Feeagh were then more irregular. Figs. 2c and 2d shows the long-term change
in the variability of the temperature measurements from Windermere and Lough
Feeagh. The points show the annual means of the standard deviations and the lines
the three-point running means. The range of variation and the pattern of change were
strikingly different at the two sites. In Lake Windermere (Fig. 2 c), the standard
deviations increased in the 1960’s and 70’s before declining in the 1980’s and 90’s.
In Lough Feeagh, the standard deviations were lower and remained more or less
constant until they increased in the 1990’s.
6
Fig. 2 The long-term trend in the surface temperature of (a) Lake Windermere and
(b) Loch Feeagh. The long-term trend in the variability (standard deviation) of the
temperature measurements from (c) Lake Windermere and (d) Lough Feeagh. The
solid lines are the three-point running means and the broken lines fitted regressions.
Factors influencing the surface temperatures and their variability
In lakes that are free of ice, the surface temperature is closely correlated with the local
air temperature.
Figs. 3a and 3b show the relationship between the surface
temperature and the air temperature at Lake Windermere and Lough Feeagh. At Lake
Windermere, there was a very strong positive correlation between the two variables (r
= 0.93, p < 0.001).
At Lough Feeagh, the correlation with the Belmullet air
temperatures was lower (r = 0.72, p < 0.001) but this site was 50 km away from the
lake. A number of factors can influence the surface temperature of a lake but the
most important are the wind speed and the cloud cover. Since there were no wind
speed measurements for the sites, our analyses were confined to cloud cover. Fig. 3c
and 3b show the relationship between the average standard deviation of the
temperature measurements at Lake Windermere and Lough Feeagh and the average
cloud cover at the meteorological stations. At Lake Windermere (Fig. 3c), there was
7
a significant negative correlation between the two variables but the variations at
Lough Feeagh were not correlated with the cloud cover at Belmullet.
Fig. 3 The relationship between the average temperature of (a) Lake Windermere and
(b) Lough Feeagh and the local air temperature. The relationship between the variability
(standard deviation) of the temperature measurements in (c) Lake Windermere and (d)
Lough Feeagh and the local cloud cover. The heavy lines show the fitted regressions and
the light lines the 95% confidence intervals.
The influence of westerly weather on the surface temperature
This weather type is typically associated with higher winds along the western
seaboard and milder, wetter conditions during the winter. Fig. 4a and 4b show the
effect that this weather type had on the surface temperatures of Lake Windermere and
Lough Feeagh. The points show the correlation between the surface temperature in a
particular week and the number of westerly days recorded over the same period. The
solid line shows the three-point running mean and the broken lines the significance
levels for the correlation coefficients. The seasonal variation in these coefficients was
very similar at the two sites. At both sites, an increase in the number of westerly days
was associated with higher temperatures in the winter and lower temperatures during
the summer.
8
Fig. 4c and 4d show the effect that westerly weather had on the variability of the
temperature measurements. In Lake Windermere (Fig. 4c), the correlations were
generally low but a few significant negative relationships were recorded during the
winter.
The pattern in Lough Feeagh (Fig. 4d) was very different.
Here, the
correlations were more variable and several significant negative relationships were
recorded during the summer. The most likely explanations for these differences are
the colder winters experienced in Windermere and the windier summers in the west of
Ireland.
During the period of study, the average winter air temperatures recorded at Ambleside
were 2.7oC lower those recorded at Bellmullet. In contrast, the average summer wind
speeds at Belmullet were 2.5 m s-1 higher than those recorded at Ambleside.
Cold winters can result in more variable surface temperatures if there are day-today variations in the rate of convective cooling. Windy summers can have the same
effect by cooling the surface or entraining colder water from deeper layers.
Fig. 4. The correlation between the weekly temperatures of (a) Lake Windermere and
(b) Lough Feeagh and the frequency of ‘westerly’ conditions. The correlation between
the variability (standard deviation) of the weekly temperatures in (c) Lake Windermere
and (d) Lough Feeagh and the frequency of ‘westerly’ conditions. The solid lines show
the three-point running means and the broken lines the significance levels.
9
THE INFLUENCE OF ANTICYCLONIC CONDITIONS ON THE
SURFACE TEMPERATURE
Anticyclonic conditions are typically associated with cold weather during the winter
and warm weather during the summer. Fig. 5a and 5b show the effect that this
weather type had on the surface temperatures of Lake Windermere and Lough
Feeagh.
The points show the correlation between the surface temperature in a
particular week and the number of anticyclonic days recorded over the same period.
The solid line shows the three-point running mean and the broken lines the
significance levels for the correlation coefficients. The seasonal variation in the
coefficients was again very similar at the two sites. At both sites, the lowest winter
and the highest summer temperatures were recorded when there was an increase in the
number of anticyclonic days. Fig. 5c and 4d show the effect that these anticyclonic
conditions had on the variability of the temperature measurements.
In Lake
Windermere (Fig. 5c), the correlations were low but a number of significant positive
correlations were recorded during the summer.
In Lough Feeagh (Fig. 5d), the
correlations were higher and a large number of ‘positive’ summer relationships were
matched by a few ‘negative’ winter relationships. These differences can again be
explained by the relative exposure of the two lakes. Since changes in cloud cover
have very little effect on the variability of the Lough Feeagh record, changes in the
wind speed must be responsible for most of the variability shown in Fig. 5d.
10
Fig. 5 The correlation between the weekly temperatures of (a) Lake Windermere and (b)
Lough Feeagh and the frequency of ‘anticyclonic’ conditions. The correlation between the
variability (standard deviation) of the weekly temperatures from (c) Lake Windermere and
(d) Lough Feeagh and the frequency of ‘anticyclonic’ conditions. The solid lines show the
three-point running means and the broken lines the significance levels.
DISCUSSION
In this paper, we have compared the thermal responses of two lakes that are exposed
to the same weather systems but are topographically rather different.
Lake
Windermere is morphometrically complex basin situated in a sheltered valley about
20 km from the sea. Lough Feeagh is a trough-like basin situated in a more open
location very close to the sea. The data assembled here showed that the surface
temperature of both lakes increased progressively during the period of study but the
rate of increase was much greater in Lake Windermere. The most likely explanation
for this difference is the more exposed situation of Lough Feeagh and its sensitivity to
wind induced mixing. The frequency of severe gales over the British Isles has
increased in recent years and there has also been an increase in the wave heights
measured in coastal waters (UKCIP, 2002). Our analyses also show that quite subtle
changes in the local weather can have a significant effect on the temperature of the
lakes. In Lake Windermere, the variability of the temperature measurements has
11
recently declined due to a progressive increase in the cloud cover. No such effect was
observed in Lough Feeagh where the variance increased over the same period.
The weather type correlations for the two lakes show that the system of weather
classification designed by Lamb for the British Isles works well for the west of
Ireland. The weekly correlations between the surface temperatures and these synoptic
categories were very similar at the two sites.
There were, however, systematic
differences in the correlations noted with the weekly variance that may be again be
related to the difference in exposure. Such results demonstrate that, lakes that are
physically very similar, can ‘filter’ the imposed climate signal in a different way. At
present, we know very little about these filtering effects.
The weather pattern
approach adopted here has the merit of quantifying these integrating effects using
indices that can be directly related to regional-scale changes in the atmospheric
circulation.
Acknowledgements
We would like to thank all the staff from the Freshwater Biological Association who
recorded the temperature measurements in Windermere. The analyses were part
funded by the European Union Environment and Climate projects REFLECT (ENV4CT97-0453) and CLIME (EVK-CT-2002-00121).
REFERENCES
George, D.G. (1989). The thermal characteristics of lakes as a measure of climate
change, Conference on Climate and Water, Helsinki, Finland. Publications of the
Academy of Finland, 402-412.
George, D.G. & Harris, G.P. (1985). The effect of climate on the long-term changes
in the crustacean zooplankton biomass of Lake Windermere, UK, Nature, 316,
536-539.
George, D.G. & Taylor, A.H. (1995). UK Lake Plankton and the Gulf Stream. Nature,
378, 139.
12
George, D.G., Talling, J.F. & Rigg, E. (2000). Factors influencing the temporal
coherence of five lakes in the English Lake District. Freshwat. Biol., 43, 449-461.
George, D.G., Maberly, S.C. & Hewitt, D.P. (2004). The influence of the North
Atlantic Oscillation on the physical, chemical and biological characteristics of four
lakes in the English Lake District. Freshwat. Biol., .
Hurrell, J.W. (1995). Decadal trends in the North Atlantic Oscillation: regional
temperature and precipitation. Science, 269, 676-679.
Kelly, P.M., Jones, P. & Briffa, K. (1997). Classifying the winds and the weather. In:
Climates of the British Isles (M.Hulme and E. Barrow eds,), Routledge, London
and New York, 454 pp.
Lamb, H.H. (1950). Types and spells of weather around the year in the British Isles.
Quart. J. Roy.Met. Soc., 76, 393-438.
Rouen, M A, George D.G & Hewitt D P. (2000). Using an automatic monitoring
station to assess the impact of episodic mixing on the seasonal succession of
phytoplankton. Verh. Internat. Verein. Limnol. 27, 2972-2976.
13